DeePM: Robust Deep Learning for Systematic Macro Portfolios

Summary

DeePM is presented as an end-to-end deep-learning system for managing diversified macro portfolios. Its design addresses asynchronous market data with a delay-based causal mechanism, constrains cross-asset relationships using an economics-informed graph, and trains against a robust risk-adjusted objective that penalizes poor historical windows as a proxy for entropic tail risk. The stated inputs are daily closing prices, and the portfolio universe consists of diversified futures.

The document reports large-scale backtests spanning 2010–2025 with transaction costs, claiming stronger net risk-adjusted performance than trend-following approaches, passive benchmarks, and a Momentum Transformer. It attributes resilience across distinct market environments to lagged cross-sectional attention, the graph prior, cost treatment, and minimax optimization. These are reported research findings rather than independently established results; the short description does not specify the exact instruments, evaluation protocol, implementation details, or uncertainty around the comparisons. Backtest performance does not establish future results.

Key ideas

• DeePM uses a delay mechanism to handle asynchronously available macro and market information.
• An economics-informed graph prior regularizes how the model represents relationships across assets.
• A worst-window penalty is used to train for robust, risk-adjusted portfolio performance.
• The reported futures backtests include transaction costs and compare the model with trend-following and passive benchmarks.
• Ablation results are said to implicate lagged attention, the graph prior, cost modeling, and robust optimization in generalization.

Tags

# DeePM: Regime-Robust Deep Learning for Systematic Macro Portfolio Management


# DeePM: Regime-Robust Deep Learning for Systematic Macro Portfolio Management









We propose DeePM (Deep Portfolio Manager), a structured deep-learning macro portfolio manager trained end-to-end to maximize a robust, risk-adjusted utility. DeePM addresses three fundamental challenges in financial learning: (1) it resolves the asynchronous "ragged filtration" problem via a Directed Delay (Causal Sieve) mechanism that prioritizes causal impulse-response learning over information freshness; (2) it combats low signal-to-noise ratios via a Macroeconomic Graph Prior, regularizing cross-asset dependence according to economic first principles; and (3) it optimizes a distributionally robust objective where a smooth worst-window penalty serves as a differentiable proxy for Entropic Value-at-Risk (EVaR) - a window-robust utility encouraging strong performance in the most adverse historical subperiods. In large-scale backtests from 2010-2025 on 50 diversified futures with highly realistic transaction costs, DeePM attains net risk-adjusted returns that are roughly twice those of classical trend-following strategies and passive benchmarks, solely using daily closing prices. Furthermore, DeePM improves upon the state-of-the-art Momentum Transformer architecture by roughly fifty percent. The model demonstrates structural resilience across the 2010s "CTA (Commodity Trading Advisor) Winter" and the post-2020 volatility regime shift, maintaining consistent performance through the pandemic, inflation shocks, and the subsequent higher-for-longer environment. Ablation studies confirm that strictly lagged cross-sectional attention, graph prior, principled treatment of transaction costs, and robust minimax optimization are the primary drivers of this generalization capability.

Shown in full with attribution under the source's licence. Licence: abstract CC0

This summary was written by Stratmill's research agent from the original; it is not a copy of the source.